The inside of the Model B after final revisions.
The wiring for the phase inverter input had to be relocated to eliminate capacitive coupling from the tone stack. It wasn’t an issue in practice except when using high-latency digital effects in the loop (think: a POD) at low master settings, in which case there would be phasing artifacts in the treble. Shortening the leads and moving them to a separate terminal strip solved the problem entirely.
Main reservoir and screen node caps had to be replaced with series-connected radials to get high-enough voltage rating for the 520V idle B+. This is only necessary with this particular PT; with a <500V B+, the supply can be greatly simplified at the cost of only ~20W of output power.
Power output statistics for this particular amp (120VAC +/-2% mains, 520V B+ idle / 482V max power, 300Hz sine test signal into resistive dummy load):
KT88: 148W fixed-screen / 135W ultralinear
6550: 145W fixed-screen / 133W ultralinear
EL34: 152W fixed-screen / 110W ultralinear
In general, I recommend against running EL34s ultralinear (poor performance), or running 6550s fixed-screen at this B+ (Vmax of 400V & Pmax of 6W for the screens is too low for long life). This amp will probably chew through tubes more quickly than most, though I knew that when I built it for maximum power. Again, lower B+ could substantially increase tube longevity and still easily exceed 100W clean, that just wasn’t my goal with this amp. I wanted to see what was possible, power-wise.
It’s a bit of a shame that frequently-revised prototypes like this end up looking so messy inside. Doing a commission build of a finalized circuit produces a much neater interior.2 notes